As companies, governments, and individuals increasingly start using AI in more and more products and processes, both to automate current activities and to develop entirely new ones, there are a few fundamental frictions that we will need to navigate: time for a thread
1) Democratization vs. centralization. AI should be a tool in the hand of everyone (like the open Internet and open personal computing), rather than being controlled by a small number of large companies and governments.
2) Human augmentation vs. replacement. Organizations should be ambitious in how they use automation, and seek to turn workers into hyper-productive "AI supervisors" (using AI to add value), rather than merely automating the work they used to be doing (using AI for cost-cutting).
3) Individual empowerment vs. mass manipulation. We should build products and systems that place the end user in control of the algorithms they interact with, rather than using AI to control people's opinions, choices, and lives. This is what I cover here: medium.com/@francois.chol…
I believe each of these 3 fundamental issues could go either way -- towards a better future or towards dystopia. That's why it is important to have these debates now. And anyone drowning out the conversation by drumming up fears of "existential risk" is doing us all a disservice
